ui-first-principles
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
bashandnodeto orchestrate its internal workflow, specifically invoking local scripts such asredact.js,normalize-input.js, andvalidate-report.js. These scripts are used for data processing, PII masking, and output validation within the project environment. - [DATA_EXFILTRATION]: No network-bound operations or data exfiltration patterns were detected. The skill operates entirely locally, processing user-provided files and scenario text on the local filesystem.
- [PROMPT_INJECTION]: The skill does not attempt to override agent instructions or bypass safety filters. Instead, it enforces strict privacy rules and provides guidance on how the agent should handle redacted content safely.
- [SAFE]: The skill demonstrates an 'assume-unsafe' security posture regarding external data. It incorporates a mandatory redaction phase (Phase 1) and a post-generation validation phase (Phase 7) that uses SHA-256 fingerprints to detect and block the leakage of original sensitive values.
Audit Metadata